因式分解串联反褶积

当前使用较多的最小相位反褶积方法均假设子波为最小相位，所以不适应非最小相位的实际地震子波，限制了地震数据处理中分辨率的提高。即使是实际地震子波为最小相位，采用最小相位反褶积方法，也会导致较多的高频噪声，使信噪比降低。本文提出的因式分解串联反褶积方法避开子波的最小相位假设，通过对自相关进行因式分解，得到子波的可能因式，然后对分解所得到的子波因式逐个求反褶积算子，并进行最大、最小相位筛选，再将各个因式串联进行试探和判别，最后实现非最小相位子波的反褶积。对理论数据和实际数据的应用表明，因式分解串联反褶积方法能显著提高分辨率，且在一定程度上提高了信噪比，即使是在最小相位子波情况下，其效果也优于当前使用的最小相位反褶积方法。

Series deconvolution by resolution into factors.

Currently more used minimum phase deconvolution method supposed that the wavelet is minimum phase, which is unsuitable for real seismic wavelet with non-minimum phase, limited the improvement of resolution in seismic data processing. Even though the real seismic wavelet is minimum phase,using minimum phase deconvolution method can also introduce more noise with high frequency,reducing S/N ratio. The method for series deconvolution by resolution into factors presented by this paper avoids the supposition that the wavelet is minimum phase. The possible factors of wavelet are obtained through resolution into factors by self-correlation,then the operator of deconvolution corresponding to each resolved factor of wavelet can be calculated separately and selection is carried out according to maximum or minimum phases, finally the try and judgment for each factor series resulted in deconvolution with non-minimum phase wavelet. The application of theoretical and real data showed that the method for series deconvolution by resolution into factors can significantly improve the resolution and S/N ratio in some degree,the effect is prior to currently-used minimum phase deconvolution method even if the wavelet is minimum phase.